October mornings in Zamagurie

I hope as we move past the first wave of AI criticism ("it doesn't work, all hype") we get a new wave of AI criticism rooted in the fact that these systems are very powerful & quite useful and focusing a deep exploration of when AI uses are uplifting and when they are detrimental

This seems…potentially good? I feel like nobody pays any attention to Microsoft AI because you can’t differentiate tools when literally everything is called Copilot

I don't wanna pile on this person bc he's just saying what everyone said 10 years ago, but I feel like when the cultural appropriation stuff came for food was when most of the reasonable people started getting annoyed lol

This is by far the best post on advanced Claude Code usage I've seen, tapping the power of hooks to context engineer and prevent slop. Only thing missing is that OP can use plugins (agents+hooks+commands+MCP) to distribute their amazing work! reddit.com/r/ClaudeAI/com…

This the hardest Halloween costume 😭 Im high as hell looking at it

No, you don’t have to make ‘halloween’ into a spooky word. It’s already spooky! The fun is shoehorning normal words into spookiness!

If a computer makes a bad management decision, it will be met with heavy gradient update Its more than can be guaranteed about many humans

this is my dream. quaint 2 bedroom house, plenty of space for the kids to run around, 20 meter radio astronomy dish

Everyone’s passwords suck. You can’t fix this. Password managers are too hard or too expensive. Passkeys are too confusing. The only thing that works is 2FA, with sms as backup

A personal favorite result: avoiding isosceles triangles on a grid. On a 64x64 grid, previous work with François Charton, Jordan Ellenberg & Geordie Williamson suggested 112 points were possible, but we couldn't find a construction. AlphaEvolve finally found this elusive set!

This is correct, and also true of every recent AI paper (the METR slow-down study, the Apple reasoning one) that casts doubt on AI's effectiveness. People are desperate to prove that LLMs don't work, aren't useful, etc. and don't really care how good the studies are.

Immense planning and technical precision was required for this absolutely preposterous (but real) view: I captured my friend Gabriel C. Brown transiting the sun during a skydive. This might be the first photo of it's kind in existence. See a video of this moment in the reply 👇

Oh thank god

If you can figure out how to configure your default state to be slightly amused rather than slightly annoyed you pretty much enter God Mode

⛄️

Gosh. This is the dream.

ChatGPT gets me

hi guys, switching things up today! Starting now, all my tweets represent the opinion of my employer

Opus 4.5 dropped while I was in the middle of something I’ve been working on for a couple of days with Sonnet. Placebos are hard, but it definitely felt like I was suddenly spending much less explaining and guiding. It just understood and got on with it.